Hi all,
How to have à resume information array about desired output Vs. obtained Output?
that means, if for example (In MNIST), for the number "2" we have 95% succees, i want to know the distribution of the remaining (5%) classifiacation error on to the remaining labels ? (On Labels "2": 80% of error was as "3", 20% error was as "8" )?
Thanks,
--